const GLboolean clamp = inst->Saturate;
const GLboolean updateCC = inst->UpdateCondRegister;
GLfloat *dstReg;
+ GLfloat dummyReg[4];
GLfloat clampedValue[4];
const GLboolean *writeMask = dest->WriteMask;
GLboolean condWriteMask[4];
case PROGRAM_TEMPORARY:
dstReg = machine->Temporaries[dest->Index];
break;
+ case PROGRAM_WRITE_ONLY:
+ dstReg = dummyReg;
+ return;
default:
_mesa_problem(NULL, "bad register file in store_vector4(fp)");
return;
/* copy existing machine */
_mesa_memcpy(dMachine, machine, sizeof(struct fp_machine));
- /* Clear temporary registers */
- _mesa_bzero( (void*) machine->Temporaries,
- MAX_NV_FRAGMENT_PROGRAM_TEMPS * 4 * sizeof(GLfloat));
+ if (program->Base.Target == GL_FRAGMENT_PROGRAM_NV) {
+ /* Clear temporary registers (undefined for ARB_f_p) */
+ _mesa_bzero( (void*) machine->Temporaries,
+ MAX_NV_FRAGMENT_PROGRAM_TEMPS * 4 * sizeof(GLfloat));
+ }
/* Add derivatives */
if (program->InputsRead & (1 << FRAG_ATTRIB_WPOS)) {
result[0] = result[1] = result[2] = result[3] =
a[0] * b[0] + a[1] * b[1] + a[2] * b[2] + a[3] * b[3];
store_vector4( inst, machine, result );
+#if DEBUG_FRAG
+ printf("DP4 %g = (%g, %g %g %g) . (%g, %g %g %g)\n",
+ result[0], a[0], a[1], a[2], a[3], b[0], b[1], b[2], b[3]);
+#endif
}
break;
case FP_OPCODE_DPH:
store_vector4( inst, machine, result );
}
break;
- case FP_OPCODE_KIL:
+ case FP_OPCODE_KIL_NV: /* NV_f_p only */
{
const GLuint *swizzle = inst->DstReg.CondSwizzle;
const GLuint condMask = inst->DstReg.CondMask;
}
}
break;
+ case FP_OPCODE_KIL: /* ARB_f_p only */
+ {
+ GLfloat a[4];
+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, a );
+ if (a[0] < 0.0F || a[1] < 0.0F || a[2] < 0.0F || a[3] < 0.0F) {
+ return GL_FALSE;
+ }
+ }
+ break;
case FP_OPCODE_LG2: /* log base 2 */
{
GLfloat a[4], result[4];
result[2] = MAX2(a[2], b[2]);
result[3] = MAX2(a[3], b[3]);
store_vector4( inst, machine, result );
+#if DEBUG_FRAG
+ printf("MAX (%g %g %g %g) = (%g %g %g %g), (%g %g %g %g)\n",
+ result[0], result[1], result[2], result[3],
+ a[0], a[1], a[2], a[3],
+ b[0], b[1], b[2], b[3]);
+#endif
}
break;
case FP_OPCODE_MIN:
GLfloat result[4];
fetch_vector4( ctx, &inst->SrcReg[0], machine, program, result );
store_vector4( inst, machine, result );
+#if DEBUG_FRAG
+ printf("MOV (%g %g %g %g)\n",
+ result[0], result[1], result[2], result[3]);
+#endif
}
break;
case FP_OPCODE_MUL:
store_vector4( inst, machine, result );
}
break;
- case FP_OPCODE_TEX:
+ case FP_OPCODE_TEX: /* Both ARB and NV frag prog */
/* Texel lookup */
{
GLfloat texcoord[4], color[4];
fetch_vector4( ctx, &inst->SrcReg[0], machine, program, texcoord );
- /* XXX: Undo perspective divide from interpolate_texcoords() */
- fetch_texel( ctx, texcoord,
- span->array->lambda[inst->TexSrcUnit][column],
- inst->TexSrcUnit, color );
+ /* Note: we pass 0 for LOD. The ARB extension requires it
+ * while the NV extension says it's implementation dependant.
+ */
+ fetch_texel( ctx, texcoord, 0.0F, inst->TexSrcUnit, color );
store_vector4( inst, machine, color );
}
break;
- case FP_OPCODE_TXB:
+ case FP_OPCODE_TXB: /* GL_ARB_fragment_program only */
/* Texel lookup with LOD bias */
{
GLfloat texcoord[4], color[4], bias, lambda;
store_vector4( inst, machine, color );
}
break;
- case FP_OPCODE_TXD:
+ case FP_OPCODE_TXD: /* GL_NV_fragment_program only */
/* Texture lookup w/ partial derivatives for LOD */
{
GLfloat texcoord[4], dtdx[4], dtdy[4], color[4];
store_vector4( inst, machine, color );
}
break;
- case FP_OPCODE_TXP:
- /* Texture lookup w/ perspective divide */
+ case FP_OPCODE_TXP: /* GL_ARB_fragment_program only */
+ /* Texture lookup w/ projective divide */
{
GLfloat texcoord[4], color[4];
fetch_vector4( ctx, &inst->SrcReg[0], machine, program, texcoord );
- /* Already did perspective divide in interpolate_texcoords() */
+ texcoord[0] /= texcoord[3];
+ texcoord[1] /= texcoord[3];
+ texcoord[2] /= texcoord[3];
+ /* Note: LOD=0 */
+ fetch_texel( ctx, texcoord, 0.0F, inst->TexSrcUnit, color );
+ store_vector4( inst, machine, color );
+ }
+ break;
+ case FP_OPCODE_TXP_NV: /* GL_NV_fragment_program only */
+ /* Texture lookup w/ projective divide */
+ {
+ GLfloat texcoord[4], color[4];
+ fetch_vector4( ctx, &inst->SrcReg[0], machine, program, texcoord );
+ if (inst->TexSrcBit != TEXTURE_CUBE_BIT) {
+ texcoord[0] /= texcoord[3];
+ texcoord[1] /= texcoord[3];
+ texcoord[2] /= texcoord[3];
+ }
fetch_texel( ctx, texcoord,
span->array->lambda[inst->TexSrcUnit][column],
inst->TexSrcUnit, color );
const GLuint *rawBits = (const GLuint *) a;
fetch_vector1( ctx, &inst->SrcReg[0], machine, program, a );
result[0] = (((rawBits[0] >> 0) & 0xff) - 128) / 127.0F;
- result[0] = (((rawBits[0] >> 8) & 0xff) - 128) / 127.0F;
- result[0] = (((rawBits[0] >> 16) & 0xff) - 128) / 127.0F;
- result[0] = (((rawBits[0] >> 24) & 0xff) - 128) / 127.0F;
+ result[1] = (((rawBits[0] >> 8) & 0xff) - 128) / 127.0F;
+ result[2] = (((rawBits[0] >> 16) & 0xff) - 128) / 127.0F;
+ result[3] = (((rawBits[0] >> 24) & 0xff) - 128) / 127.0F;
store_vector4( inst, machine, result );
}
break;
const GLuint *rawBits = (const GLuint *) a;
fetch_vector1( ctx, &inst->SrcReg[0], machine, program, a );
result[0] = ((rawBits[0] >> 0) & 0xff) / 255.0F;
- result[0] = ((rawBits[0] >> 8) & 0xff) / 255.0F;
- result[0] = ((rawBits[0] >> 16) & 0xff) / 255.0F;
- result[0] = ((rawBits[0] >> 24) & 0xff) / 255.0F;
+ result[1] = ((rawBits[0] >> 8) & 0xff) / 255.0F;
+ result[2] = ((rawBits[0] >> 16) & 0xff) / 255.0F;
+ result[3] = ((rawBits[0] >> 24) & 0xff) / 255.0F;
store_vector4( inst, machine, result );
}
break;
if (ctx->FragmentProgram.CallbackEnabled)
inputsRead = ~0;
- /* Clear temporary registers */
- _mesa_bzero(machine->Temporaries,
- MAX_NV_FRAGMENT_PROGRAM_TEMPS * 4 * sizeof(GLfloat));
+ if (program->Base.Target == GL_FRAGMENT_PROGRAM_NV) {
+ /* Clear temporary registers (undefined for ARB_f_p) */
+ _mesa_bzero(machine->Temporaries,
+ MAX_NV_FRAGMENT_PROGRAM_TEMPS * 4 * sizeof(GLfloat));
+ }
/* Load input registers */
if (inputsRead & (1 << FRAG_ATTRIB_WPOS)) {
init_machine(ctx, &ctx->FragmentProgram.Machine,
ctx->FragmentProgram.Current, span, i);
+#ifdef USE_TCC
+ if (!_swrast_execute_codegen_program(ctx, program, ~0,
+ &ctx->FragmentProgram.Machine,
+ span, i)) {
+ span->array->mask[i] = GL_FALSE; /* killed fragment */
+ }
+#else
if (!execute_program(ctx, program, ~0,
&ctx->FragmentProgram.Machine, span, i)) {
span->array->mask[i] = GL_FALSE; /* killed fragment */
}
+#endif
/* Store output registers */
{